EP1043875B1 - Empfänger für diskrete Mehrträger-modulierte Signale mit Fensterfunktion - Google Patents

Empfänger für diskrete Mehrträger-modulierte Signale mit Fensterfunktion Download PDF

Info

Publication number
EP1043875B1
EP1043875B1 EP99122054A EP99122054A EP1043875B1 EP 1043875 B1 EP1043875 B1 EP 1043875B1 EP 99122054 A EP99122054 A EP 99122054A EP 99122054 A EP99122054 A EP 99122054A EP 1043875 B1 EP1043875 B1 EP 1043875B1
Authority
EP
European Patent Office
Prior art keywords
stage
output
receiver
signal
window
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
EP99122054A
Other languages
English (en)
French (fr)
Other versions
EP1043875A3 (de
EP1043875A2 (de
Inventor
Samir Kapoor
Slobodan Nedic
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NEC USA Inc
Original Assignee
NEC USA Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by NEC USA Inc filed Critical NEC USA Inc
Publication of EP1043875A2 publication Critical patent/EP1043875A2/de
Publication of EP1043875A3 publication Critical patent/EP1043875A3/de
Application granted granted Critical
Publication of EP1043875B1 publication Critical patent/EP1043875B1/de
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L27/00—Modulated-carrier systems
    • H04L27/26—Systems using multi-frequency codes
    • H04L27/2601—Multicarrier modulation systems
    • H04L27/2647—Arrangements specific to the receiver only

Definitions

  • the present invention relates to a receiver for DMT based (Discrete Multitone Modulation) signals and preferably for DMT ADSL (Asymmetric Digital Subscriber Line) and Lite ADSL modems.
  • DMT ADSL Asymmetric Digital Subscriber Line
  • Lite ADSL modems DMT ADSL
  • windowing and allied techniques Some related techniques have been previously considered as possible solutions to combat narrow band interference (NBI) and spectrally colored crosstalk for DMT transceivers.
  • Windowing or more generally, pulse/waveform shaping filtering, can be carried out over single or multiple DMT symbols and can be carried out jointly or singly at the transmitter and receiver.
  • Several existing methods require the participation of the transmitter and/or utilize window/pulse shapes that span one or more DMT symbols and improve performance based on better spectral containment.
  • any methods which require the transmitter to perform waveform shaping will either have to be standardized or will result in a proprietary standard non-compliant modem.
  • There exist methods which do not require transmitter participation but do require substantially equalized channels e.g. NEC's ADSL modem has provision for partial shaping at the boundaries of the orthogonality interval).
  • the present invention will be described in the context of a wireline communications channel, such as a telephone line which utilizes a twisted pair of copper wires. It is noted that the use of the present invention is not limited to wireline systems as those skilled in the art will appreciate from the discussion hereinbelow.
  • a modem is typically used to transmit and receive digital data over a telephone line.
  • Modems employ a modulator to transmit the digital data over the telephone line and a demodulator to receive digital data from the telephone line.
  • One common modulation technique is known as discrete multi-tone modulation (DMT) which requires a discrete multi-tone transmitter and a discrete multi-tone receiver at each modem in a communication system.
  • DMT discrete multi-tone modulation
  • Fig. 14 is a block diagram of a conventional DMT communications system 1.
  • the system 1 includes a DMT transmitter 10, a transmission channel 20, and a DMT receiver 30.
  • the DMT transmitter 10 includes a symbol generator 12, an inverse fast Fourier transform (IFFT) modulator 14 and a cyclic prefix generator 16.
  • the DMT transmitter 10 receives an input bit stream b(n) which is fed into the symbol generator 12.
  • the symbol generator 12 produces a signal X(k) which is fed into the IFFT modulator 14.
  • X(k) is a complex signal (i.e., a signal understood by those skilled in the art to comprise both a real and an imaginary component) formed by mapping groups of bits of the input bit stream b(n) into a complex data space such that the complex signal X(k) has a length of N samples.
  • Symbol generator 12 also augments the signal X(k) with a complex conjugate to obtain a conjugate symmetric signal of 2N samples.
  • the IFFT modulator 14 performs a 2N-point inverse fast fourier transform on the conjugate complex signal X(k) to obtain the sampled real signal x(n). Since X(k) is a symmetric signal, the output of the IFFT modulator 16 is a real signal x(n).
  • the real signal x(n) may be thought of as the summation of a plurality of cosine functions each having a finite length and a different frequency, phase, and amplitude, where these frequencies are multiples of a fundamental frequency. Since each of the cosine functions has a finite duration, x(n) is a varying amplitude discrete signal having a finite duration spanning 2N samples. Each cosine function is known as a bin or tone.
  • the transmission channel 20 is modeled as including a D/A converter 22, transmit filter (not shown), a receive filter (not shown), and an A/D converter 26 on either end of a wire loop 24.
  • D/A converter 22 and the transmit filter
  • receive filter not shown
  • A/D converter 26 and the receive filter
  • the frequency spectrum of x(n) may be thought of as a plurality of orthogonal (SIN X)/(X) functions, each centered at a respective one of the frequencies of the cosine functions of x(n).
  • x(n) is transmitted over the channel 20 to che DMT receiver 30. Since the transmission channel 20 has a non-ideal impulse response h(n), the received signal y(n) will not exactly match x(n). Instead, y(n) will be a function of the convolution of x(n) and h(n) . Typically, h(n) will look substantially like the curve shown in Fig. 15. The non-ideal characteristic of h(n) introduces an amount of interference (specifically intersymbol and interchannel interference) which should be compensated for in both the DMT transmitter 10 and the DMT receiver 30.
  • interference specifically intersymbol and interchannel interference
  • a common technique in compensating for the non-ideal impulse response of the transmission channel 20 is to introduce a so-called guard band at the beginning of each finite duration signal x(n) to produce x'(n).
  • the cyclic prefix generator 16 performs this function.
  • the guard band is typically formed of the last G samples of x(n) for each DMT symbol. If the length of the impulse response h(n) of the transmission channel 20 is less than or equal to G + 1, then the guard band of length G will be sufficient to eliminate the interference caused by the impulse response h(n).
  • the guard band is commonly referred to in the art as a "cyclic prefix" (CP).
  • the impulse response h(n) of a typical transmission channel 20 may be excessively long, requiring cyclic prefix lengths which substantially reduce the rate at which digital bits are transmitted across the transmission channel 20.
  • the DMT receiver 30, therefore, employs signal processing techniques which effectively shorten the impulse response h(n) of the transmission channel 20, thereby permitting a corresponding reduction in the length of the cyclic prefix required at the DMT transmitter 10.
  • the DMT receiver 30 includes a time-domain equalizer (TEQ) 32, CP remover 34 for removing the cyclic prefix, a fast fourier transform (FFT) demodulator 36, and a bit generator 38.
  • the time-domain equalizer 32 is a finite impulse response (FIR) filter designed to compensate for the non-ideal impulse response h(n) of the transmission channel 20.
  • the time-domain equalizer 32 employs a finite number of coefficients (T) which are calculated to compensate for the non-ideal impulse response of the transmission channel 20.
  • the time domain equalizer 32 operates on the impulse response h(n) of the channel 20 such that the combined impulse response h eff (n) of the channel 20 and the time domain equalizer 32 has maximum energy within a limited band of samples. This may be thought of as "shortening" the effective impulse response of the channel 20.
  • the output of the time domain equalizer is z'(n).
  • the CP remover 34 is employed to remove the cyclic prefix from z'(n) to obtain z(n).
  • the signal z(n) is input into the FFT demodulator 36 (which is understood to include a one-tap per-bin frequency domain equalizer/AGC function) to produce the complex symmetric signal X(k).
  • the bit generator 38 maps the complex signal X(k) into an output bit stream b(n), which theoretically matches the input bit stream b(n).
  • the conventional DMT receiver 30 of Fig. 14 operates optimally when white noise is present at the output of the time domain equalizer 32, it is susceptible to increased interference when colored noise is present. This is particularly pronounced when che colored noise exhibits spectral nulls or spectral peaks.
  • Colored noise may be present at the output of the time domain equalizer 32 because (i) additive colored noise was injected into the signal x'(n) as it was transmitted over the transmission channel 20; and/or (ii) the time domain equalizer 32 itself introduces spectral shaping (especially spectral nulls/peaks) into the signal z'(n). Thus, even if the transmission channel 20 does not introduce additive colored noise into the received signal y(n), the time domain equalizer 32 may itself introduce speccral coloration into the additive noise of signal z'(n).
  • time domain equalizer 32 may produce a "shorter" effective impulse response h eff (n), it may degrade system performance by introducing colored noise (especially spectral nulls/peaks) into z'(n).
  • colored noise especially spectral nulls/peaks
  • the rate at which data bits b(n) are transmitted over the transmission channel 20 and the error rate of such transmission may be adversely affected by colored noise at the output of the time domain equalizer 32.
  • an improved DMT communication system which is capable of (i) improved performance in the presence of noise, and in particular, narrowband interference and inter-bin interference; (ii) compensating for additive colored noise and/or narrowband interference (NBI) introduced by the transmission channel and/or other conditions leading to diminished orthogonality between bins including inadequate channel shortening, symbol timing offset and jitter; (iii) mitigating against the spectral coloration of additive noise by the time domain equalizer; (iv) suppressing side lobes caused by DFT frequency response; (v) obtaining better performance against cross-talk and narrowband interference (NBI) without changing the transmitter; and/or (vi) suppressing a specific type colored noise in the form of local echo signal cross-talk in the FDM duplexing method
  • EP 802 649 A discloses a discrete multi-tone receiver comprising in sequence a time domain equalizer, a window function unit, a discrete Fourier transformer and a frequency domain equalizer
  • the windowing function is given a particular shape to reduce spectral leakage from one frequency to another and thereby reduce the influence of band limited noise on the DMT carriers.
  • the present invention provides a receiver for DMT-ADSL modems which does not require transmitter pulse shaping.
  • the receiver of the invention carries out windowing independently for each symbol over its entire orthogonality interval (excluding the cyclic prefix). Either one of two equivalent methods may be used for windowing - time-domain windowing (TDW) or frequency-domain windowing (FDW).
  • TDW time-domain windowing
  • FDW frequency-domain windowing
  • a conventional demodulator consisting of an FFT (Fast Fourier Transform) type detector followed by a single-tap-per-bin frequency-domain equalizer (FEQ) is augmented by an additional data-path utilizing windowing.
  • a decision feedback equalizer (DFE) type detector is used at the output of the windowed data-path to cancel inter-bin-interference (IBI) created by windowing.
  • IBI inter-bin-interference
  • Pilot-bins are employed to reduce the threat of error propagation in the DFE.
  • the receiver architecture of the invention does not require any changes to the ADSL standard specifications (such as ANSI T1.413 issue 2, ITU-T G.992.1 and ITU-T G.992.2) and is relatively simple to implement.
  • NBI narrowband interference
  • Limited performance gains may also be achieved in the presence of crosstalk with strong spectral coloration.
  • the present invention does not require any cooperation from the transmitter, rendering it impervious to standardization issues (at least in the case of ADSL standards).
  • the overall method of the invention is relatively simple to implement and can provide significant performance gains for certain interference conditions compared to other windowing and equalization techniques.
  • DMT Downlink Traffic Management
  • VDSL Very High Bit-Rate Digital Subscriber Line
  • wireless OFDM Orthogonal Frequency Division Multiplexing
  • Windowing according to the invention can be carried out in the time-domain (TDW) or in the frequency-domain (FDW).
  • TDW time-domain
  • FDW frequency-domain
  • TDW is done by a sample-by-sample multiplication of the received signal samples by the window coefficients prior to the FFT in a second data path.
  • FDW is done after the FFT on a bin-by-bin basis by taking a linear combination of bin outputs and frequency-domain window coefficients.
  • a conventional demodulator is provided in a first data path.
  • the conventional demodulator may use a single-tap-per-bin FEQ for each bin after the FFT.
  • the second daca path is the windowed data-path and uses a DFE to cancel the IBI caused by windowing.
  • the DFE makes use of prior decisions made on sub-symbols (bins) within the same DMT symbol.
  • An alternative implementation of the FDW can also be carried out by using the output of the FEQ in the first data path instead of the FFT output. This also necessitates a change in the DFE coefficients, though its overall structure is unchanged.
  • pilot bins (deterministic sub-symbols) can be utilized.
  • the receiver has a bin-select logic stage which selects the output of either data path for input to a slicer. This determination can be made after the achieved SNR (Signal to Noise Ratio) in each bin has been measured from each data path and corresponding bit-loading profiles have been computed.
  • a DFE error-propagation SNR penalty is applied to the achieved SNR in the windowed data path before computing its bit-loading profile.
  • One criterion to select either data path for a given bin is to select the path which yields higher bits-per-bin (the unwindowed data path can be chosen in case of a tie) . This issue is discussed in more detail below.
  • a DMT signal model will be discussed.
  • N orthogonal sinusoids tones/bins
  • I i and Q i are the real and imaginary parts of the ith complex frequency domain sub-symbol
  • a i I i + jQ i chosen from a QAM constellation
  • j ⁇ -1
  • CP cyclic prefix
  • Typical values for b min and b max are 6 and 31 respectively (implementation dependent).
  • the CP merely serves as a guard time , albeit with a special cyclical structure. Inter-symbol-interference is eliminated and orthogonality between bins is preserved when the CP duration exceeds the channel impulse response (CIR) i.e. when G ⁇ U.
  • CIR channel impulse response
  • the effective discrete-time CIR encountered by the above signal include the analog communication channel, transmitter D/A, receiver A/D, interpolation and front-end filtering in the ATU-C and ATU-R.
  • CP remover 34 See Fig.
  • a conventional receiver performs a Discrete Fourier Transform (DFT) operation via the FFT algorithm each symbol time. See FFT demodulator 36 of Fig. 14. This is followed by a single-tap-per-bin frequency domain equalizer (FEQ) contained within FFT demodulator 36 of Fig. 14.
  • DFT Discrete Fourier Transform
  • a key aspect of the inventive technique which differentiates it from the conventional configuration (FFT followed by FEQ) described above is the use of a parallel windowing stage followed by a decision feedback equalizer (DFE) to cancel the window-induced inter-bin-interference (IBI). Windowing can be implemented equivalently in the time- or frequency-domains depending on the window type, computational constraints and convenience of implementation. See Figs. 1 and 2.
  • Fig. 1 shows a TDW implementation and Fig. 2 shows an FDW implementation.
  • Figure 1 shows a DMT receiver using time domain windowing.
  • the invention is employed after the CP remover 34 of Fig. 14,i.e., the input is provided from the output of CP remover 34, preferably after being time domain equalized by TEQ 32.
  • the output of the CP remover is provided to two data paths 100 and 200.
  • the first data path 100 (DP1) is a conventional data path employing FFT 120 and an FEQ 140.
  • the output of the FEQ is provided to the bin select logic 300.
  • time domain windowing is performed by the time domain window 220.
  • the output of the time domain window 220 is provided to an FFT 240 whose output is provided into the DFE 260.
  • the output of FFT 120 is also provided to the DFE 260 via line 122.
  • the output of DFE 260 is provided to the bin select logic 300, which will be described in more detail below.
  • the output of the bin select logic 300 is provided to a slicer circuit 400 which provides the demodulated output.
  • the demodulated output is also provided in the feedback path 410 to the DFE 260.
  • FIG. 2 shows the DMT demodulator employing frequency domain windowing.
  • the CP remover 34 output is provided to a first data path 100' (DP1) which includes the FFT 120' and an FEQ 140'.
  • the output of the FFT 120' is provided to a second data path 200' (DP2) where frequency domain windowing is performed.
  • the output of the frequency domain window 220' is provided to the DFE 260'.
  • the output of the DFE 260' is provided to the bin select logic 300'.
  • the output of FEQ 140' from the first data path 100' is also provided to the bin select logic.
  • the output of the bin select logic 300' is provided to slicer circuit 400' which provides the output of the demodulator.
  • a feedback path 410' is provided from the output of the slicer circuit 400' to the DFE 260'.
  • FEQ 140' provides coefficients for DFE 260' via path 142'.
  • the goal of receiver windowing is to enable suppression of the side-lobes of the Discrete Fourier Transform (DFT) frequency response and consequently obtain better performance against crosstalk and narrowband interference (NBI), while keeping the transmitter unchanged.
  • DFT Discrete Fourier Transform
  • Time-domain windowing is performed by a sample-by-sample multiplication of 2N received signal samples each symbol time with the window coefficients.
  • TDW Time-domain windowing
  • a variety of windows may be used.
  • FIG. 3 depicts the frequency response of several windows.
  • the DPS Discrete Prolate Spheroidal window is designed to have the narrowest main-lobe whose side-lobes are no higher than -50 dB relative to the main-lobe peak.
  • the upper box 100 marked as Data-Path 1 (DP1) is the conventional demodulator described above.
  • the lower box 200 marked as Data-Path 2 (DP2) is the new portion including time domain windowing.
  • Incoming samples simultaneously pass through both DP1 and DP2.
  • the ith demodulated sub-symbol (after FFT) in DP1 is denoted by â i and in DP2 by b and i . Then, where
  • ⁇ i denotes the window-induced IBI experienced by the ith bin from all other active bins.
  • the multiplication of the window samples with the DMT symbol samples in the time-domain followed by an FFT is. equivalent to a frequency-domain convolution.
  • windowing can be carried out efficiently in the frequency domain as
  • Equations (2.3) and (2.4) suggest that windowing can be looked upon as a form of partial response signaling (PRS) along frequency with coefficients c 1 .
  • PRS partial response signaling
  • the Hanning window can be effectively viewed as a class V PRS response.
  • An important difference between conventional PRS and receiver windowing as in the present invention is that the PRS function is carried out at the transmitter while windowing according to the invention is carried out at the receiver-end including the effect of channel attenuation coefficients and additive noise.
  • Training is a two stage-process:
  • This structure bears similarity with DFE structures used in a classical PRS system. This is not surprising because of the basic mathematical similarity in terms of time-frequency duality in the two approaches, notwithstanding the different motivations behind these techniques.
  • Single-carrier PRS equalizers/detectors operate in the time-domain with symbol-spaced taps while the DFE here operates in the frequency-domain with bin-spaced taps.
  • a key difference is that the effect of the channel ( ⁇ i ) is taken into account in the DFE here while this is not an issue in PRS equalizers.
  • the Hanning window displays the best side-lobe suppression.
  • Fig. 5A shows the alternative demodulator architecture with frequency domain windowing.
  • the output of the CP remover 34 is provided to Fast Fourier Transform 120''.
  • the output of the FFT 120'' in first data path 100'' is provided to the frequency domain equalizer (FEQ) 140''.
  • the output of the FEQ 140'' represent the equalized subsymbols .
  • the equalized subsymbols are provided to the bin select logic stage 300'' whose output is provided to the slicer 400''.
  • the equalized subsymbols are also provided into the second data path 200'' comprising the frequency domain window stage 220'' and the decision feedback equalizer (DFE) 260''.
  • DFE decision feedback equalizer
  • the equalized subsymbols output from the DFE 260'' are provided to the bin select logic 300''.
  • the bin select logic selects the output from either the first path 100'' or the second path 200'' depending upon set criterion, e.g., which path produces more data bits per symbol or has a higher SNR.
  • the output from the slicer 400'' is provided in a feedback path to the DFE 260'' to cancel interference created by the frequency domain window stage 220''.
  • ⁇ i elements in x i right are obtained as reconstructed sub-symbol values obtained from the slicer. Care must be taken to normalize the slicer 400, 400', 400'' output by taking the ADSL gain scaling factors and constellation energy normalization scaling factors into account before substituting them back into x i .
  • the bin-select logic 300, 300', 300'' determines the data-path (DP1 or DP2) used to provide the slicer input for any given sub-symbol.( Figure 4 shows slicer operating on ).
  • the bin-select logic may select all sub-symbols from DP1 only (i.e., no windowing is used), all from DP2 only, or select from DP1 or DP2 on a bin by bin basis. For example, in a simulation described below, a simple criterion of selecting the path which yields higher number of bits in a given bin is used. Once a SNR-per-bin profile is computed for DP1 and DP2, the bit-loading algorithm is run twice - once on each profile from DP1 and DP2. This ensures that the basic bit-loading algorithm being used need not be modified and that the combined bit-allocation is no worse than a conventional receiver having path DP1 only.
  • a symbol-rate LMS update of FEQ (and thus, the DFE) weights can be carried out to track variations in the channel impulse response (CIR) and interference.
  • Sub-rate adaptation may also be used for economy of implementation. Normalized LMS is preferred over regular LMS to make the adaptation speed impervious to the dynamic range of the input data for all bins.
  • > d min, i 2 )] 2 ⁇ 4(1 - 1 M ) Q ( d min, i 2 ⁇ U ) where the Q-function is defined as and d min,i
  • d denotes the minimum distance between QAM constellation points at the channel output and d equals the distance between uncoded input QAM constellation points.
  • the probability of symbol error per dimension (Pe/2) is generally specified (le-7 for ADSL/ADSL-Lite).
  • the upper bound increases exponentially with the number of feedback taps.
  • the BER increases are large when there are multiple feedback coefficients (large L), the feedback coefficients have relatively large values (c 1 /c 0 ratios are not small) and when multiple bits are loaded per sub-symbol (large M).
  • adjacent sub-symbols in DMT can belong to different sized constellations, i.e. different values of M for adjacent (feedback) bins. This further complicates a potential analytical solution for equation (2.15).
  • DMT symbols enjoy the benefit of deterministic pilot sub-symbols each symbol time, e.g. bin 64 downstream and bin 16 upstream in T1.413).
  • precoding to counter error propagation is well known in the literature. However, for the problem at hand, this approach suffers from some drawbacks. First, it requires participation of the transmitter. This precludes an independent application of the overall technique unless standardization considerations are met. Second, precoding does not necessarily provide immunity against error propagation depending on the constellation size and feedback coefficients. In fact, it has been shown to even marginally degrade performance for some cases. Third, the precoder is likely to be far more complex than the conventional precoders used in PRS systems. This is because the effect of channel coefficients and unequal QAM constellations in adjacent feedback bins has to be explicitly taken into account.
  • Test cases are presented in Figures 7 - 13 for downstream transmission.
  • Figure 6 depicts the two example loops used for the plots.
  • the first loop (ANSI T1.601 loop 2) is 16.5 Kft with one bridge tap.
  • the second loop (CSA Loop 4) is a medium length 7.6 Kft loop with 2 bridge taps.
  • a Least Squares based channel shortening TEQ algorithm with channel estimation was used in all cases to shorten the CIR except in Figure 13.
  • Channel estimation was performed using impulse trains in the presence of the impairments such as crosstalk or NBI. With 544 samples/symbol, the 10-tap FIR TEQ filter requires 5,440 real multiplications and 4896 real additions per DMT symbol time.
  • a 2-tap FDW (Window 1) and a 3-tap Hanning window (Window 2) are used for all cases shown with bin 64 (pilot) as the reference bin.
  • bin 63 is also used as a reference to initiate the equalization process.
  • the DFE input vector is X i right while X i left is used for the others.
  • Window 1 is seen to provide a good trade-off between having fewer feedback taps, good side-lobe suppression and low SNR degradation in white noise.
  • the spikes seen in the bit-allocation plots at bin 64 are due to the pilot bin which is always modulated with a constant 2-bit sub-symbol.
  • the gain-scaling factor on alternate bins is set to zero at the transmitter. This ensures that there is no IBI between adjacent bins and only a single-tap FEQ (as in DP1) is required. This, of course, also reduces the number of active bins by half but can serve as a good indication of what is potentially achievable through windowing.
  • the signal and noise (crosstalk or NBI) PSD's are plotted before and after the channel shortening TEQ.
  • AWGN of PSD -140 dBm/Hz is added to the crosstalk and NBI for all cases. It can be seen that the TEQ alters the spectrum of the signal and noise at the input of DP1 and DP2 and thus influences the performance of the rest of the receiver.
  • the achieved SNR per bin is plotted for the conventional method (identical to DP1) and with windowing (DP2).
  • the resulting bit-loading tables depict the resulting bit loading profiles for le-7 BER with 4 dB margin for DP1 and using Window 1 (bit-loading arising from Window 2 is not plotted).
  • Bit-loading for DP2 also includes an additional SNR margin of 1.14 dB as described above.
  • the bin-select logic for selecting DP1 or DP2 for any given bin is simply to select the path which yields higher bits per bin after bit-loading is complete. In case of a tie, DP1 is chosen.
  • the combined (maximum) bit allocation is marked by "circles”.
  • Example cases plotted in Figures 7- 13. FIGURE Loop Crosstalk NBI TEQ 7, 8 Long -- Yes Yes 9, 10 Medium -- Yes Yes 11, 12 Medium 10 ADSL-US -- Yes 13 Medium 10 ADSL-US -- -- --
  • Figures 7 and 8 show that the DFE provides the most gain in the vicinity of the NBI.
  • Window 2 performs better than Window 1 in the immediate vicinity of the NBI but incurs a greater loss in the rest of the band.
  • the effect of NBI monotonically reduces with frequency separation. Thus, bins spaced further in frequency are less affected and the performance advantage of windowing gradually reduces until it approaches the white-noise loss of 3 dB for Window 1.
  • Figures 9 and 10 display similar performance for the medium loop. It is notable that performance gains are dependent on the strength of the NBI relative to the ambient SNR determined by the received signal power after passing through the channel and the AWGN floor. In Figure 9, the effect of the NBI is more pronounced compared to Figure 7 because the ambient SNR is higher. Thus, NBI suppression is more effective and greater gains are obtained via windowing.
  • Figures 11 and 12 show that the DFE provides no advantage in the presence of crosstalk which is not strongly colored after the TEQ (Time Domain Equalizer).
  • TEQ Time Domain Equalizer
  • Figure 13 is an example of the dramatic improvement that can be obtained by superior spectral confinement. This case is the same as Figure 11(A) above except that no TEQ is used.
  • the CIR is not confined to the CP resulting in additional noise due to IBI and ISI (Inter Symbol Interference). Since the per-bin SNR due to AWGN and crosstalk is relatively much higher, the effect of IBI+ISI is isolated.
  • the achieved SNR per bin in DP1 (solid curve) essentially represents the noise floor due to channel induced IBI+ISI. Windowing with subsequent use of DFE substantially improves the per-bin SNR in this case resulting in an approximately three-fold increase in the aggregate data rate. Similar results have been obtained for other loops as well. In cases when the TEQ is present, but does not adequately shorten the CIR, corresponding performance gains are obtained. Imperfect symbol alignment and timing jitter also lead to diminished orthogonality between bins and can benefit from this approach. Preliminary results also suggest that the invention can provide some additional margin against local transmitter echo (which can effectively appear as a strong spectrally colored cross-talker).
  • an overall SNR enhancement in any particular situation depends on the relative levels of the ambient SNR due to additive noise and the IBI+ISI noise due to imperfect channel shortening. Since only the IBI+ISI component is suppressed via windowing, significant (if any) improvement will be obtained only when the ambient SNR is relatively higher.
  • the TDW operation requires 2N real multiplications when a 2N-pt FFT is used while the DFE complexity is the same as above. However, most of the additional complexity lies in doing an extra FFT. Optimized implementations of FFT's often incur fewer multiplications/additions than shown above by exploiting the fact that the input is purely real and only the first N outputs are required (output is conjugate symmetric).
  • TDW and FDW are equivalent, there can be a significant difference in computational load. All windows with real coefficients result in equal TDW complexity.
  • the computational load in FDW depends on the number of feedback coefficients, i.e. on the particular FDW frequency response.
  • FDW is independent of the preceding FFT size and can be carried out on a bin-by-bin basis as required. Provision for implementing multiple windows can be done by making the number of FDW coefficients and their values programmable.
  • different bins can have different windows applied to them (in the same symbol) if desired, (including no windowing).
  • the window coefficients can be programmable as well and FFT blocks used in other algorithms (such as for the TEQ or for PAR (Peak-to-Average Ratio) reduction) may be available for reuse.
  • Alternative embodiments of TDW can potentially be used to reduce the implementation complexity in terms of hardware and/or required instructions. For instance, the two FFT's in Figure 1 can be executed by the same hardware/firmware block in a sequential fashion (staggered in time) instead of being done simultaneously by two different blocks.
  • 2-tap FDWs (1 feedback tap) are well-suited since they provide a good balance between side-lobe suppression and complexity.
  • Windows with multiple feedback taps suffer from the drawbacks of increased risk of error propagation, need for greater number of reference sub-symbols and greater DFE computation.
  • the SNR degradation in white noise only is also greater.
  • DPS windows are a class of windows which can be tailored for different side-lobe suppression versus main-lobe width trade-off at the expense of greater number of DFE taps ( Figure 3) and can be used in some situations.
  • the FEQ and DFE can adapt to NBI and/or crosstalk if it emerges sufficiently slowly relative to the tracking capabilities of the LMS (Least Mean Squares) algorithm and change the bit-allocation via bit-swaps. More elaborate and potentially faster tracking RLS (Recursive Least Squares) algorithms may also be used. If a strong interferer appears suddenly, the modem may fall back into fast retrain (in ADSL-Lite) or even a full restart. In any case, the sustainable data rate (or SNR margin) after the incident is likely to be better, or equivalently require fewer bit-swaps.
  • LMS Local Mean Squares
  • RLS Recursive Least Squares
  • SNR Margin SNR Margin or data-rate. This follows from the fact that, depending on the operative constraints, a higher achievable SNR in bins is equivalent to (A) Greater loop reach for a given minimum margin and data rate or (B) Higher SNR margin in active data carrying bins for a given data-rate and loop reach or (C) Greater data-rate for a given reach and margin.
  • Gains in SNR using the invention are not unconditional, i.e ., gains may be realized only in the presence of NBI, highly correlated crosstalk, or other imperfections such as inadequate channel shortening, symbol alignment offsets and jitter.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
  • Digital Transmission Methods That Use Modulated Carrier Waves (AREA)

Claims (46)

  1. Empfänger zum Bereitstellen eines demodulierten Ausgangssignals (ã) von einem empfangenen diskreten Multitonmodulationseingangssignal, wobei das Eingangssignal von einem verrauschten Kommunikationskanal empfangen wird, das Eingangssignal einen darauf aufmodulierten digitalen Bitstrom aufweist, und wobei der Empfänger aufweist:
    einen ersten (100, 100', 100'') und einen zweiten (200, 200', 200'') Datenpfad, die so verbunden sind, daß sie das Eingangssignal empfangen;
       wobei der erste der beiden Datenpfade (100, 100', 100'') eine erste Stufe mit einer Frequenzantwort zum Anwenden einer diskreten Fouriertransformation (120, 120', 120'') auf das Eingangssignal und ferner einen Frequenzbereichsentzerrer (140, 140', 140'') mit einem mit einem Ausgang der ersten Stufe verbundenen Eingang aufweist;
       wobei der zweite der beiden Datenpfade (200, 200', 200'') eine Fensterstufe (220, 220', 220'') zum Unterdrücken von Nebenkeulen der Frequenzantwort der ersten Stufe aufweist; und
       eine Logikstufe (300, 300', 300'') zum Auswählen eines Ausgangssignals vom ersten Datenpfad oder vom zweiten Datenpfad basierend auf einem vordefinierten Test und zum Bereitstellen eines den demodulierten digitalen Bitstrom darstellenden ausgewählten Ausgangssignals.
  2. Empfänger nach Anspruch 1, ferner mit einem im zweiten Datenpfad angeordneten Entscheidungs-Feedback-Entzerrer (260, 260', 260''), dem als Eingangssignal ein Ausgangssignal der Fensterstufe zugeführt wird, zum Unterdrücken von durch die Fensterstufe erzeugter Inter-Bin-Interferenz.
  3. Empfänger nach Anspruch 2, wobei der Entscheidungs-Feedback-Entzerrer ein Rückkopplungssignal (410, 410', 410'') von einem Ausgang der Logikstufe empfängt, um Inter-Bin-Interferenz zu unterdrücken.
  4. Empfänger nach Anspruch 3, ferner mit einer Slicer-Stufe (400, 400', 400''), der als Eingangssignal das Ausgangssignal der Logikstufe zugeführt wird, zum Bereitstellen des als das Rückkopplungssignal bereitgestellten demodulierten digitalen Bitstroms (ã) als Ausgangssignal.
  5. Empfänger nach Anspruch 1, 2, 3 oder 4, wobei die Fensterstufe im zweiten Datenpfad eine Zeitbereichfensterstufe (220) ist.
  6. Empfänger nach Anspruch 5, wobei der Zeitbereichfensterstufe ein Eingangssignal direkt von einem Zeitbereichsentzerrer zugeführt wird.
  7. Empfänger nach Anspruch 6, ferner mit einer zweiten Stufe zum Ausführen einer diskreten Fouriertransformation (240) im zweiten Datenpfad, wobei die zweite Stufe einen Eingang aufweist, an dem ein Ausgangssignal von der Zeitbereichfensterstufe empfangen wird.
  8. Empfänger nach Anspruch 7, ferner mit einem im zweiten Datenpfad angeordneten Entscheidungs-Feedback-Entzerrer (260), dem ein Eingangssignal von der zweiten Stufe zugeführt wird, und dessen Ausgangssignal der Logikstufe zugeführt wird.
  9. Empfänger nach Anspruch 8, wobei der Entscheidungs-Feedback-Entzerrer ein weiteres Eingangssignal (122) von der ersten Stufe empfängt.
  10. Empfänger nach einem der Ansprüche 1 bis 4, wobei die Fensterstufe im zweiten Datenpfad eine Frequenzbereichfensterstufe (220', 220'') ist.
  11. Empfänger nach Anspruch 10, ferner mit einem im zweiten Datenpfad angeordneten Entscheidungs-Feedback-Entzerrer (260', 260'') mit einem Eingang, dem ein Ausgangssignal von der Frequenzbereichfensterstufe zugeführt wird, wobei der Entscheidungs-Feedback-Entzerrer zum Unterdrücken von durch die Frequenzbereichfensterstufe erzeugter Inter-Bin-Interferenz dient, und wobei ein Ausgangssignal des Entscheidungs-Feedback-Entzerrers der Logikstufe zugeführt wird.
  12. Empfänger nach Anspruch 11, wobei dem Entscheidungs-Feedback-Entzerrer (260') ferner ein weiteres Eingangssignal (142') von der ersten Stufe zugeführt wird.
  13. Empfänger nach Anspruch 12, wobei dem Entscheidungs-Feedback-Entzerrer ein Rückkopplungssignal (410') zum Unterdrücken von durch die Frequenzbereichfensterstufe erzeugter Inter-Bin-Interferenz von einem Ausgang der Logikstufe zugeführt wird.
  14. Empfänger nach Anspruch 13, wobei die Logikstufe (300') einer Slicer-Stufe (400') ein Ausgangssignal zuführt, wobei die Slicer-Stufe den demodulierten digitalen Bitstrom als Ausgangssignal (ã) bereitstellt, und wobei das Rückkopplungssignal das demodulierte Ausgangssignal enthält.
  15. Empfänger nach Anspruch 10, 11, 12, 13 oder 14, wobei die Frequenzbereichfensterstufe einen mit einem Ausgang der ersten Stufe verbundenen Eingang aufweist.
  16. Empfänger nach einem der Ansprüche 1 bis 15, wobei die Logikstufe (300, 300', 300'') ein Ausgangssignal vom ersten Datenpfad oder vom zweiten Datenpfad basierend auf einer Entscheidung darüber auswählt, welcher Datenpfad mehr Bits pro Symbol oder einen höheren Rauschabstand bereitstellt.
  17. Empfänger nach Anspruch 16, wobei die Logikstufe im Fall einer Verbindung den ersten Datenpfad auswählt.
  18. Empfänger nach einem der Ansprüche 1 bis 17, wobei die Fensterstufe eine Hanning-Fensterfunktion, eine DPS-Fensterfunktion oder eine Bartlett-Fensterfunktion aufweist.
  19. Empfänger nach einem der Ansprüche 5 bis 9, wobei die Zeitbereichfensterstufe (220) eine Zeitbereichpulsformung durch abtastwertweises Multiplizieren des Ausgangssignals vom Zeitbereichsentzerrer mit Fensterkoeffizienten ausführt, die eine Zeitbereichfensterfunktion der Zeitbereichfensterstufe definieren.
  20. Empfänger nach einem der Ansprüche 10 bis 15, wobei die Frequenbereichfensterstufe eine Frequenzbereichpulsformung auf einer bin-weisen Basis durch Ausführen einer Linearkombination von Bin-Ausgangssignalen und Frequenzbereichfensterkoeffizienten ausführt, die eine Frequenzbereichfensterfunktion der Frequenzbereichfensterstufe definieren.
  21. Empfänger nach einem der Ansprüche 1 bis 20, wobei der Frequenzbereichsentzerrer (140, 140', 140'') ein Entzerrer mit einem Abgriff pro Bin ist.
  22. Empfänger nach einem der Ansprüche 1 bis 21, ferner mit einem Zeitbereichsentzerrer (32), dem als Eingangssignal das vom Kommunikationskanal empfangene Signal zugeführt wird, wobei der Zeitbereichsentzerrer ein Ausgangssignal erzeugt, das dem ersten und dem zweiten Datenpfad zugeführt wird.
  23. Empfänger nach einem der Ansprüche 10 bis 15, wobei der Fensterstufe (220'') als Eingangssignal ein Ausgangssignal des Frequenzbereichsentzerrers (140'') zugeführt wird.
  24. Verfahren zum Bereitstellen eines demodulierten Ausgangssignals von einem empfangenen diskreten Multitonmodulationseingangssignal, wobei das Eingangssignal von einem verrauschten Kommunikationskanal empfangen wird, wobei das Eingangssignal einen darauf aufmodulierten digitalen Bitstrom aufweist, wobei das Verfahren die Schritte aufweist:
    Zuführen des Eingangssignals zum ersten (100, 100', 100'') und zweiten (200, 200', 200'') Datenpfad;
    Anwenden einer diskreten Fouriertransformation (120, 120', 120'') auf das Eingangssignal im ersten Datenpfad, um ein erstes transformiertes Signal (142') zu erzeugen, und Frequenzbereichsentzerren (140, 140', 140'') des ersten transformierten Signals, um ein frequenzbereichsentzerrtes Signal bereitzustellen;
    Unterdrücken von Nebenkeulen der Frequenzantwort des ersten transformierten Signals im zweiten Datenpfad durch Anwenden einer Fensterfunktion (220, 220', 220''), um ein pulsgeformtes Signal bereitzustellen; und
    Auswählen (300, 300', 300'') eines Ausgangssignals vom ersten Datenpfad oder vom zweiten Datenpfad basierend auf einem vordefinierten Test und Bereitstellen eines den demodulierten digitalen Bitstrom darstellenden ausgewählten Ausgangssignals.
  25. Verfahren nach Anspruch 24, ferner mit dem Schritt zum Ausführen einer Entscheidungs-Feedback-Entzerrung (260, 260', 260'') im zweiten Datenpfad bezüglich des pulsgeformten Signals zum Unterdrücken von durch die Fensterfunktion erzeugter Inter-Bin-Interferenz.
  26. Verfahren nach Anspruch 25, ferner mit einem während des Entscheidungs-Feedback-Entzerrungsschritts ausgeführten Schritt zum Bereitstellen eines Rückkopplungssignals (410, 410', 410''), das das ausgewählte Ausgangssignal enthält, zum Unterdrücken von Inter-Bin-Interferenz.
  27. Verfahren nach Anspruch 26, ferner mit den Schritten zum Teilen (400, 400', 400'') des ausgewählten Ausgangssignals und zum Bereitstellen des geteilten Ausgangssignals als demodulierter digitaler Bitstrom, wobei der demodulierte digitale Bitstrom als Rückkopplungssignal bereitgestellt wird.
  28. Verfahren nach Anspruch 24, 25, 26 oder 27, wobei der Schritt zum Anwenden einer Fensterfunktion im zweiten Datenpfad das Anwenden einer Zeitbereichfensterfunktion (220) aufweist.
  29. Verfahren nach Anspruch 28, wobei der Schritt zum Anwenden einer Zeitbereichfensterfunktion das direkte Anwenden einer Zeitbereichfensterfunktion auf das Eingangssignal aufweist.
  30. Verfahren nach Anspruch 29, ferner mit einem nach dem Schritt zum Anwenden einer Zeitbereichfensterfunktion ausgeführten Schritt zum Ausführen einer diskreten Fouriertransformation (240) im zweiten Datenpfad, um ein zweites transformiertes Signal (b and) bereitzustellen.
  31. Verfahren nach Anspruch 30, ferner mit einem Schritt zum Ausführen einer Entscheidungs-Feedback-Entzerrung (260) des zweiten transformierten Signals im zweiten Datenpfad.
  32. Verfahren nach Anspruch 31, ferner mit dem Schritt zum Verwenden des ersten transformierten Signals (142') während des Entscheidungs-Feedback-Entzerrungsschritts (260).
  33. Verfahren nach einem der Ansprüche 24 bis 27, wobei der Schritt zum Anwenden einer Fensterfunktion im zweiten Datenpfad das Anwenden einer Frequenzbereichfensterfunktion (220', 220'') aufweist.
  34. Verfahren nach Anspruch 33, ferner mit den Schritten zum Ausführen einer Entscheidungs-Feedback-Entzerrung (260', 260'') im zweiten Datenpfad bezüglich des pulsgeformten Signals, wobei der Entscheidungs-Feedback-Entzerungsschritt zum Unterdrücken von durch die Frequenzbereichfensterfunktion erzeugter Inter-Bin-Interferenz bereitgestellt wird, und Bereitstellen eines als Ausgangssignal auszuwählenden Entscheidungs-Feedback-entzerrten Signals (
    Figure 00550001
    ).
  35. Verfahren nach Anspruch 34, ferner mit dem Schritt zum Verwenden des ersten transformierten Signals (142') während des Entscheidungs-Feedback-Entzerrungsschritts (260'').
  36. Verfahren nach Anspruch 35, ferner mit dem Schritt zum Bereitstellen eines Rückkopplungssignals (410') während des Entscheidungs-Feedback-Entzerrungsschritts zum Unterdrücken der durch die Frequenzbereichfensterfunktion erzeugten Inter-Bin-Interferenz.
  37. Verfahren nach Anspruch 36, ferner mit dem Schritt zum Teilen (400) des ausgewählten Ausgangssignals in den demodulierten digitalen Bitstrom (ã), wobei das Rückkopplungssignal den demodulierten digitalen Bitstrom enthält.
  38. Verfahren nach einem der Ansprüche 33 bis 37, ferner mit dem Schritt zum Anwenden der Frequenzbereichfensterfunktion (220') auf das erste transformierte Signal.
  39. Verfahren nach einem der Ansprüche 24 bis 38, wobei der Schritt (300, 300', 300'') zum Auswählen eines Ausgangssignals vom ersten Datenpfad oder vom zweiten Datenpfad das Auswählen eines Ausgangssignals basierend auf einer Entscheidung aufweist, gemäß der entschieden wird, welcher Datenpfad mehr Bits pro Symbol bereitstellt oder einen größeren Rauschabstand aufweist.
  40. Verfahren nach Anspruch 39, wobei im Fall einer Blockierung der Schritt zum Auswählen eines Ausgangssignals das Auswählen des ersten Datenpfades aufweist.
  41. Verfahren nach einem der Ansprüche 24 bis 40, wobei die Fensterfunktion eine Hanning-Fensterfunktion, eine DPS-Fensterfunktion oder eine Bartlett-Fensterfunktion aufweist.
  42. Verfahren nach einem der Ansprüche 28 bis 32, wobei der Schritt zum Anwenden einer Zeitbereichfensterfunktion eine Zeitbereichpulsformung durch abtastwertweises Multiplizieren des zeitbereichsentzerrten Signals mit Fensterkoeffizienten aufweist, die eine Zeitbereichfensterfunktion definieren.
  43. Verfahren nach einem der Ansprüche 33 bis 38, wobei der Schritt zum Anwenden einer Frequenzbereichfensterfunktion eine Frequenzbereichpulsformung auf einer bin-weisen Basis durch Ausführen einer Linearkombination von Bin-Ausgangssignalen und Frequenzbereichfensterkoeffizienten aufweist, die eine Frequenzbereichfensterfunktion definieren.
  44. Verfahren nach einem der Ansprüche 24 bis 38, wobei der Frequenzbereichsentzerrungsschritt (140, 140', 140'') das Frequenzbereichsentzerren unter Verwendung eines Entzerrers mit einem Abgriff pro Bin aufweist.
  45. Verfahren nach einem der Ansprüche 24 bis 44, ferner mit den Schritten zum Zeitbereichsentzerren (32) des vom Kommunikationskanal empfangenen Eingangssignals zum Erzeugen eines zeitbereichsentzerrten Signals und zum Zuführen des zeitbereichsentzerrten Signals zum ersten und zum zweiten Datenpfad.
  46. Verfahren nach einem der Ansprüche 33 bis 38, ferner mit dem Schritt zum Anwenden der Fensterfunktion (220'') auf das frequenzbereichsentzerrte Signal (
    Figure 00570001
    ).
EP99122054A 1999-04-08 1999-11-17 Empfänger für diskrete Mehrträger-modulierte Signale mit Fensterfunktion Expired - Lifetime EP1043875B1 (de)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US09/288,478 US7023938B1 (en) 1999-04-08 1999-04-08 Receiver for discrete multitone modulated signals having window function
US288478 1999-04-08

Publications (3)

Publication Number Publication Date
EP1043875A2 EP1043875A2 (de) 2000-10-11
EP1043875A3 EP1043875A3 (de) 2002-04-03
EP1043875B1 true EP1043875B1 (de) 2004-07-28

Family

ID=23107276

Family Applications (1)

Application Number Title Priority Date Filing Date
EP99122054A Expired - Lifetime EP1043875B1 (de) 1999-04-08 1999-11-17 Empfänger für diskrete Mehrträger-modulierte Signale mit Fensterfunktion

Country Status (4)

Country Link
US (1) US7023938B1 (de)
EP (1) EP1043875B1 (de)
JP (1) JP2000307481A (de)
DE (1) DE69918945T2 (de)

Families Citing this family (37)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
SE9801748L (sv) * 1998-05-18 1999-11-19 Telia Ab Förbättringar i eller som hänför sig till telekommunikationsöverföringssystem
US6563841B1 (en) * 1999-08-30 2003-05-13 Nec Usa, Inc. Per-bin adaptive equalization in windowed DMT-type modem receiver
EP1157512A1 (de) 1999-12-30 2001-11-28 Bandspeed, Inc. Verfahren zur verarbeitung von einem kommunikationskanal empfangenen daten bei anwendungen begrenzter arithmetischer präzision
DE60127395T2 (de) 2000-04-18 2007-12-06 Aware, Inc., Bedford Datenzuweisung mit änderbaren signal-rauschabstand
US6996199B2 (en) * 2001-01-25 2006-02-07 Bandspeed, Inc. Approach for processing data received from a communications channel to reduce noise power and optimize impulse response length to reduce inter-symbol interference and inter-channel interference
US7272199B2 (en) * 2001-01-25 2007-09-18 Bandspeed, Inc. Adaptive adjustment of time and frequency domain equalizers in communications systems
US7023937B2 (en) * 2001-03-01 2006-04-04 Texas Instruments Incorporated Receiver window design for multicarrier communication systems
US6856649B2 (en) * 2001-03-30 2005-02-15 Koninklijke Philips Electronics N.V. Initialization scheme for a hybrid frequency-time domain equalizer
US7042937B2 (en) 2001-04-23 2006-05-09 Koninklijke Philips Electronics N.V. Hybrid frequency-time domain equalizer
US7274736B2 (en) * 2001-10-24 2007-09-25 Texas Instruments Incorporated Multiple path equalization for multicarrier systems
US7263119B1 (en) 2001-11-29 2007-08-28 Marvell International Ltd. Decoding method and apparatus
DE10160733A1 (de) * 2001-12-11 2003-08-28 Siemens Ag Verfahren zur kontinuierlichen Schätzung der Entzerrkoeffizienten für leitungsgebundene Übertragungssysteme
US7139343B2 (en) * 2002-03-27 2006-11-21 Hitachi Kokusai Electric Inc. Digital radio receiver
WO2003090422A1 (en) * 2002-04-19 2003-10-30 Conexant Systems, Inc. System and method for detecting narrowband interference in a wideband ofdm sign al based on normalized subcarrier amplitudes
WO2003098853A1 (fr) * 2002-05-17 2003-11-27 Matsushita Electric Industrial Co., Ltd. Dispositif de reception, procede de reception, et dispositif de mesure des caracteristiques d'un canal de transmission
US7453949B2 (en) 2003-12-09 2008-11-18 Agere Systems Inc. MIMO receivers having one or more additional receive paths
JP4199144B2 (ja) * 2004-03-11 2008-12-17 株式会社東芝 ウェイト関数生成装置、参照信号生成装置、送信信号生成装置、信号処理装置及びアンテナ装置
US7477683B2 (en) * 2004-03-29 2009-01-13 Stmicroelectronics Ltd. Periodic DMT signals with cyclic extension
WO2006069475A1 (fr) * 2004-12-28 2006-07-06 Zte Corporation Procede permettant d'eliminer l'interference entre porteuses dans un systeme de communication mobile a multiplexage par repartition orthogonale de la frequence
JP4675790B2 (ja) * 2006-01-27 2011-04-27 三菱電機株式会社 通信装置および通信システム
KR101200736B1 (ko) * 2006-03-14 2012-11-13 퀄컴 인코포레이티드 주파수 영역의 신호 처리를 위한 스마트 안테나 시스템 및이를 위한 이동 단말과 기지국의 송수신 장치 및 방법
HUE045130T2 (hu) * 2006-09-15 2019-12-30 Ikanos Communications Inc Eljárás és eszköz különbözõ robusztusságú kommunikációs csatornákhoz többhangú adó-vevõben
GB0707355D0 (en) * 2007-04-16 2007-05-23 Nxp Bv Channel estimation
US8321491B2 (en) * 2007-06-26 2012-11-27 The United States Of America As Represented By The Secretary Of The Army System and method for detecting a weak signal in a noisy environment
US8321823B2 (en) * 2007-10-04 2012-11-27 Carnegie Mellon University System and method for designing architecture for specified permutation and datapath circuits for permutation
JP4875642B2 (ja) * 2008-02-27 2012-02-15 株式会社日立製作所 雑音電力推定装置及び方法
US8295395B2 (en) * 2008-09-30 2012-10-23 Apple Inc. Methods and apparatus for partial interference reduction within wireless networks
TWI416955B (zh) * 2008-11-24 2013-11-21 Realtek Semiconductor Corp 單載波／多載波共用接收器以及處理單載波/多載波共用接收器所接收之接收訊號的方法
CN101808059B (zh) * 2009-02-13 2013-04-24 瑞昱半导体股份有限公司 单载波/多载波共享接收器
EP2228955B1 (de) * 2009-03-13 2017-02-22 OCT Circuit Technologies International Limited Vorrichtung und Verfahren zum Empfang von OFDM-Signalen basierend auf Zeitbereichsfensterung in Anwesenheit vom Dopplereffekt
US8755460B2 (en) * 2010-07-30 2014-06-17 National Instruments Corporation Phase aligned sampling of multiple data channels using a successive approximation register converter
US8804809B2 (en) 2011-09-12 2014-08-12 Transwitch Corporation Techniques for setting feedback coefficients of a PAM-N decision feedback equalizer
US8908796B1 (en) * 2013-05-15 2014-12-09 University Of South Florida Orthogonal frequency division multiplexing (OFDM) transmitter and receiver windowing for adjacent channel interference (ACI) suppression and rejection
US8971396B1 (en) 2013-08-22 2015-03-03 Pmc-Sierra Us, Inc. Windowed-based decision feedback equalizer and decision feedback sequence estimator
US10282387B2 (en) * 2013-11-06 2019-05-07 Nxp Usa, Inc. FFT device and method for performing a Fast Fourier Transform
JP6935370B2 (ja) * 2018-07-24 2021-09-15 アンリツ株式会社 信号発生装置および該装置を用いた周波数特性表示方法
CN109871575B (zh) * 2018-12-29 2022-12-20 陕西海泰电子有限责任公司 一种基于时域fft的电磁干扰接收机窗函数的设计方法

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100331437B1 (ko) * 1995-06-30 2002-08-08 삼성전자 주식회사 디.엠.티.시스템에서적응형비트교환방법및장치
FR2743967B1 (fr) * 1996-01-18 1998-03-27 France Telecom Procede et dispositif de synchronisation temporelle d'un recepteur d'un signal multiporteuse
EP0802649A1 (de) * 1996-04-19 1997-10-22 ALCATEL BELL Naamloze Vennootschap Verfahren und Fensteranordnung zum Reduzieren von Verlusten, Fourier-Transformation und DMT-Modem wobei diese Anordnung verwendet wird
US6014412A (en) * 1996-04-19 2000-01-11 Amati Communications Corporation Digital radio frequency interference canceller
US6353629B1 (en) * 1997-05-12 2002-03-05 Texas Instruments Incorporated Poly-path time domain equalization
US6047025A (en) * 1998-02-13 2000-04-04 Motorola, Inc. Method and apparatus for equallization in an asymmetric digital aubscriber line communications system
US6396886B1 (en) * 1999-02-12 2002-05-28 Nec Usa, Inc. DMT time-domain equalizer algorithm

Also Published As

Publication number Publication date
EP1043875A3 (de) 2002-04-03
US7023938B1 (en) 2006-04-04
EP1043875A2 (de) 2000-10-11
DE69918945T2 (de) 2004-12-16
JP2000307481A (ja) 2000-11-02
DE69918945D1 (de) 2004-09-02

Similar Documents

Publication Publication Date Title
EP1043875B1 (de) Empfänger für diskrete Mehrträger-modulierte Signale mit Fensterfunktion
Benvenuto et al. On the comparison between OFDM and single carrier modulation with a DFE using a frequency-domain feedforward filter
US6266367B1 (en) Combined echo canceller and time domain equalizer
US6097763A (en) MMSE equalizers for DMT systems with cross talk
US6549512B2 (en) MDSL DMT architecture
EP0933897B1 (de) Verringerung von Interferenz in Mehrtonübertragungssystemen mit Duplexbetrieb
US8265215B1 (en) Method and system for determining symbol boundary timing in a multicarrier data transmission system
JP2007195218A (ja) デジタル無線周波数干渉打消器
US6452987B1 (en) Fast start-up in discrete multi-tone (DMT) based communications system
US8170125B2 (en) Subdimensional single-carrier modulation
JPH11196060A (ja) マルチバンド検出器
KR100441250B1 (ko) 이퀄라이저의 계수 계산 방법 및 그것을 계산하는 장치
US7817730B2 (en) Training sequence for symbol boundary detection in a multicarrier data transmission system
US6563841B1 (en) Per-bin adaptive equalization in windowed DMT-type modem receiver
US7289554B2 (en) Method and apparatus for channel equalization and cyclostationary interference rejection for ADSL-DMT modems
US8385399B2 (en) Method for impulse noise mitigation
US20050111538A1 (en) Method and apparatus for improving the quality of channel estimation algorithms using training sequences
Kapoor et al. Interference suppression in DMT receivers using windowing
US20110228838A1 (en) Apparatus for impulse noise mitigation
Milosevic et al. DMT bit rate maximization with optimal time domain equalizer filter bank architecture
Zeng et al. Near-end crosstalk mitigation in ADSL systems
US7274736B2 (en) Multiple path equalization for multicarrier systems
Shim et al. Complexity analysis of multicarrier and single-carrier systems for very high-speed digital subscriber line
Khan et al. Channel equalization for Discrete Wavelet Multitone transceiver in wireline channels
Münster et al. MMSE Channel Prediction for Symbol-by-symbol Adaptive OFDM Systems

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

Kind code of ref document: A2

Designated state(s): DE GB

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Kind code of ref document: A3

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

RIC1 Information provided on ipc code assigned before grant

Free format text: 7H 04L 27/26 A, 7H 04L 1/00 B

17P Request for examination filed

Effective date: 20020220

AKX Designation fees paid

Free format text: DE GB

GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE GB

REG Reference to a national code

Ref country code: GB

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: IE

Ref legal event code: FG4D

REF Corresponds to:

Ref document number: 69918945

Country of ref document: DE

Date of ref document: 20040902

Kind code of ref document: P

REG Reference to a national code

Ref country code: GB

Ref legal event code: 711B

REG Reference to a national code

Ref country code: GB

Ref legal event code: 711G

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed

Effective date: 20050429

REG Reference to a national code

Ref country code: GB

Ref legal event code: 732E

Free format text: REGISTERED BETWEEN 20090917 AND 20090923

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20091112

Year of fee payment: 11

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20091111

Year of fee payment: 11

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20101117

REG Reference to a national code

Ref country code: DE

Ref legal event code: R119

Ref document number: 69918945

Country of ref document: DE

Effective date: 20110601

Ref country code: DE

Ref legal event code: R119

Ref document number: 69918945

Country of ref document: DE

Effective date: 20110531

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20110531

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20101117